[PATCH] pacct: add pacct_struct to fix some pacct bugs.
[linux-2.6-block.git] / include / linux / sched.h
CommitLineData
1da177e4
LT
1#ifndef _LINUX_SCHED_H
2#define _LINUX_SCHED_H
3
b7b3c76a
DW
4#include <linux/auxvec.h> /* For AT_VECTOR_SIZE */
5
6/*
7 * cloning flags:
8 */
9#define CSIGNAL 0x000000ff /* signal mask to be sent at exit */
10#define CLONE_VM 0x00000100 /* set if VM shared between processes */
11#define CLONE_FS 0x00000200 /* set if fs info shared between processes */
12#define CLONE_FILES 0x00000400 /* set if open files shared between processes */
13#define CLONE_SIGHAND 0x00000800 /* set if signal handlers and blocked signals shared */
14#define CLONE_PTRACE 0x00002000 /* set if we want to let tracing continue on the child too */
15#define CLONE_VFORK 0x00004000 /* set if the parent wants the child to wake it up on mm_release */
16#define CLONE_PARENT 0x00008000 /* set if we want to have the same parent as the cloner */
17#define CLONE_THREAD 0x00010000 /* Same thread group? */
18#define CLONE_NEWNS 0x00020000 /* New namespace group? */
19#define CLONE_SYSVSEM 0x00040000 /* share system V SEM_UNDO semantics */
20#define CLONE_SETTLS 0x00080000 /* create a new TLS for the child */
21#define CLONE_PARENT_SETTID 0x00100000 /* set the TID in the parent */
22#define CLONE_CHILD_CLEARTID 0x00200000 /* clear the TID in the child */
23#define CLONE_DETACHED 0x00400000 /* Unused, ignored */
24#define CLONE_UNTRACED 0x00800000 /* set if the tracing process can't force CLONE_PTRACE on this clone */
25#define CLONE_CHILD_SETTID 0x01000000 /* set the TID in the child */
26#define CLONE_STOPPED 0x02000000 /* Start in stopped state */
27
28/*
29 * Scheduling policies
30 */
31#define SCHED_NORMAL 0
32#define SCHED_FIFO 1
33#define SCHED_RR 2
34#define SCHED_BATCH 3
35
a3b6714e 36#ifdef __KERNEL__
b7b3c76a
DW
37
38struct sched_param {
39 int sched_priority;
40};
41
1da177e4
LT
42#include <asm/param.h> /* for HZ */
43
1da177e4
LT
44#include <linux/capability.h>
45#include <linux/threads.h>
46#include <linux/kernel.h>
47#include <linux/types.h>
48#include <linux/timex.h>
49#include <linux/jiffies.h>
50#include <linux/rbtree.h>
51#include <linux/thread_info.h>
52#include <linux/cpumask.h>
53#include <linux/errno.h>
54#include <linux/nodemask.h>
55
56#include <asm/system.h>
57#include <asm/semaphore.h>
58#include <asm/page.h>
59#include <asm/ptrace.h>
60#include <asm/mmu.h>
61#include <asm/cputime.h>
62
63#include <linux/smp.h>
64#include <linux/sem.h>
65#include <linux/signal.h>
66#include <linux/securebits.h>
67#include <linux/fs_struct.h>
68#include <linux/compiler.h>
69#include <linux/completion.h>
70#include <linux/pid.h>
71#include <linux/percpu.h>
72#include <linux/topology.h>
73#include <linux/seccomp.h>
e56d0903 74#include <linux/rcupdate.h>
0771dfef 75#include <linux/futex.h>
1da177e4 76
a3b6714e
DW
77#include <linux/time.h>
78#include <linux/param.h>
79#include <linux/resource.h>
80#include <linux/timer.h>
81#include <linux/hrtimer.h>
82
83#include <asm/processor.h>
36d57ac4 84
1da177e4
LT
85struct exec_domain;
86
1da177e4
LT
87/*
88 * List of flags we want to share for kernel threads,
89 * if only because they are not used by them anyway.
90 */
91#define CLONE_KERNEL (CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
92
93/*
94 * These are the constant used to fake the fixed-point load-average
95 * counting. Some notes:
96 * - 11 bit fractions expand to 22 bits by the multiplies: this gives
97 * a load-average precision of 10 bits integer + 11 bits fractional
98 * - if you want to count load-averages more often, you need more
99 * precision, or rounding will get you. With 2-second counting freq,
100 * the EXP_n values would be 1981, 2034 and 2043 if still using only
101 * 11 bit fractions.
102 */
103extern unsigned long avenrun[]; /* Load averages */
104
105#define FSHIFT 11 /* nr of bits of precision */
106#define FIXED_1 (1<<FSHIFT) /* 1.0 as fixed-point */
107#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
108#define EXP_1 1884 /* 1/exp(5sec/1min) as fixed-point */
109#define EXP_5 2014 /* 1/exp(5sec/5min) */
110#define EXP_15 2037 /* 1/exp(5sec/15min) */
111
112#define CALC_LOAD(load,exp,n) \
113 load *= exp; \
114 load += n*(FIXED_1-exp); \
115 load >>= FSHIFT;
116
117extern unsigned long total_forks;
118extern int nr_threads;
119extern int last_pid;
120DECLARE_PER_CPU(unsigned long, process_counts);
121extern int nr_processes(void);
122extern unsigned long nr_running(void);
123extern unsigned long nr_uninterruptible(void);
db1b1fef 124extern unsigned long nr_active(void);
1da177e4
LT
125extern unsigned long nr_iowait(void);
126
1da177e4 127
4a8342d2
LT
128/*
129 * Task state bitmask. NOTE! These bits are also
130 * encoded in fs/proc/array.c: get_task_state().
131 *
132 * We have two separate sets of flags: task->state
133 * is about runnability, while task->exit_state are
134 * about the task exiting. Confusing, but this way
135 * modifying one set can't modify the other one by
136 * mistake.
137 */
1da177e4
LT
138#define TASK_RUNNING 0
139#define TASK_INTERRUPTIBLE 1
140#define TASK_UNINTERRUPTIBLE 2
4a8342d2
LT
141#define TASK_STOPPED 4
142#define TASK_TRACED 8
143/* in tsk->exit_state */
144#define EXIT_ZOMBIE 16
145#define EXIT_DEAD 32
146/* in tsk->state again */
147#define TASK_NONINTERACTIVE 64
1da177e4
LT
148
149#define __set_task_state(tsk, state_value) \
150 do { (tsk)->state = (state_value); } while (0)
151#define set_task_state(tsk, state_value) \
152 set_mb((tsk)->state, (state_value))
153
498d0c57
AM
154/*
155 * set_current_state() includes a barrier so that the write of current->state
156 * is correctly serialised wrt the caller's subsequent test of whether to
157 * actually sleep:
158 *
159 * set_current_state(TASK_UNINTERRUPTIBLE);
160 * if (do_i_need_to_sleep())
161 * schedule();
162 *
163 * If the caller does not need such serialisation then use __set_current_state()
164 */
1da177e4
LT
165#define __set_current_state(state_value) \
166 do { current->state = (state_value); } while (0)
167#define set_current_state(state_value) \
168 set_mb(current->state, (state_value))
169
170/* Task command name length */
171#define TASK_COMM_LEN 16
172
1da177e4
LT
173#include <linux/spinlock.h>
174
175/*
176 * This serializes "schedule()" and also protects
177 * the run-queue from deletions/modifications (but
178 * _adding_ to the beginning of the run-queue has
179 * a separate lock).
180 */
181extern rwlock_t tasklist_lock;
182extern spinlock_t mmlist_lock;
183
184typedef struct task_struct task_t;
185
186extern void sched_init(void);
187extern void sched_init_smp(void);
188extern void init_idle(task_t *idle, int cpu);
189
190extern cpumask_t nohz_cpu_mask;
191
192extern void show_state(void);
193extern void show_regs(struct pt_regs *);
194
195/*
196 * TASK is a pointer to the task whose backtrace we want to see (or NULL for current
197 * task), SP is the stack pointer of the first frame that should be shown in the back
198 * trace (or NULL if the entire call-chain of the task should be shown).
199 */
200extern void show_stack(struct task_struct *task, unsigned long *sp);
201
202void io_schedule(void);
203long io_schedule_timeout(long timeout);
204
205extern void cpu_init (void);
206extern void trap_init(void);
207extern void update_process_times(int user);
208extern void scheduler_tick(void);
209
8446f1d3 210#ifdef CONFIG_DETECT_SOFTLOCKUP
6687a97d 211extern void softlockup_tick(void);
8446f1d3
IM
212extern void spawn_softlockup_task(void);
213extern void touch_softlockup_watchdog(void);
214#else
6687a97d 215static inline void softlockup_tick(void)
8446f1d3
IM
216{
217}
218static inline void spawn_softlockup_task(void)
219{
220}
221static inline void touch_softlockup_watchdog(void)
222{
223}
224#endif
225
226
1da177e4
LT
227/* Attach to any functions which should be ignored in wchan output. */
228#define __sched __attribute__((__section__(".sched.text")))
229/* Is this address in the __sched functions? */
230extern int in_sched_functions(unsigned long addr);
231
232#define MAX_SCHEDULE_TIMEOUT LONG_MAX
233extern signed long FASTCALL(schedule_timeout(signed long timeout));
64ed93a2
NA
234extern signed long schedule_timeout_interruptible(signed long timeout);
235extern signed long schedule_timeout_uninterruptible(signed long timeout);
1da177e4
LT
236asmlinkage void schedule(void);
237
238struct namespace;
239
240/* Maximum number of active map areas.. This is a random (large) number */
241#define DEFAULT_MAX_MAP_COUNT 65536
242
243extern int sysctl_max_map_count;
244
245#include <linux/aio.h>
246
247extern unsigned long
248arch_get_unmapped_area(struct file *, unsigned long, unsigned long,
249 unsigned long, unsigned long);
250extern unsigned long
251arch_get_unmapped_area_topdown(struct file *filp, unsigned long addr,
252 unsigned long len, unsigned long pgoff,
253 unsigned long flags);
1363c3cd
WW
254extern void arch_unmap_area(struct mm_struct *, unsigned long);
255extern void arch_unmap_area_topdown(struct mm_struct *, unsigned long);
1da177e4 256
f412ac08
HD
257#if NR_CPUS >= CONFIG_SPLIT_PTLOCK_CPUS
258/*
259 * The mm counters are not protected by its page_table_lock,
260 * so must be incremented atomically.
261 */
d3cb4871
CL
262#define set_mm_counter(mm, member, value) atomic_long_set(&(mm)->_##member, value)
263#define get_mm_counter(mm, member) ((unsigned long)atomic_long_read(&(mm)->_##member))
264#define add_mm_counter(mm, member, value) atomic_long_add(value, &(mm)->_##member)
265#define inc_mm_counter(mm, member) atomic_long_inc(&(mm)->_##member)
266#define dec_mm_counter(mm, member) atomic_long_dec(&(mm)->_##member)
267typedef atomic_long_t mm_counter_t;
f412ac08
HD
268
269#else /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
270/*
271 * The mm counters are protected by its page_table_lock,
272 * so can be incremented directly.
273 */
1da177e4
LT
274#define set_mm_counter(mm, member, value) (mm)->_##member = (value)
275#define get_mm_counter(mm, member) ((mm)->_##member)
276#define add_mm_counter(mm, member, value) (mm)->_##member += (value)
277#define inc_mm_counter(mm, member) (mm)->_##member++
278#define dec_mm_counter(mm, member) (mm)->_##member--
f412ac08
HD
279typedef unsigned long mm_counter_t;
280
281#endif /* NR_CPUS < CONFIG_SPLIT_PTLOCK_CPUS */
4294621f 282
f412ac08
HD
283#define get_mm_rss(mm) \
284 (get_mm_counter(mm, file_rss) + get_mm_counter(mm, anon_rss))
365e9c87
HD
285#define update_hiwater_rss(mm) do { \
286 unsigned long _rss = get_mm_rss(mm); \
287 if ((mm)->hiwater_rss < _rss) \
288 (mm)->hiwater_rss = _rss; \
289} while (0)
290#define update_hiwater_vm(mm) do { \
291 if ((mm)->hiwater_vm < (mm)->total_vm) \
292 (mm)->hiwater_vm = (mm)->total_vm; \
293} while (0)
294
1da177e4
LT
295struct mm_struct {
296 struct vm_area_struct * mmap; /* list of VMAs */
297 struct rb_root mm_rb;
298 struct vm_area_struct * mmap_cache; /* last find_vma result */
299 unsigned long (*get_unmapped_area) (struct file *filp,
300 unsigned long addr, unsigned long len,
301 unsigned long pgoff, unsigned long flags);
1363c3cd 302 void (*unmap_area) (struct mm_struct *mm, unsigned long addr);
0551fbd2
BH
303 unsigned long mmap_base; /* base of mmap area */
304 unsigned long task_size; /* size of task vm space */
305 unsigned long cached_hole_size; /* if non-zero, the largest hole below free_area_cache */
1363c3cd 306 unsigned long free_area_cache; /* first hole of size cached_hole_size or larger */
1da177e4
LT
307 pgd_t * pgd;
308 atomic_t mm_users; /* How many users with user space? */
309 atomic_t mm_count; /* How many references to "struct mm_struct" (users count as 1) */
310 int map_count; /* number of VMAs */
311 struct rw_semaphore mmap_sem;
312 spinlock_t page_table_lock; /* Protects page tables and some counters */
313
314 struct list_head mmlist; /* List of maybe swapped mm's. These are globally strung
315 * together off init_mm.mmlist, and are protected
316 * by mmlist_lock
317 */
318
f412ac08
HD
319 /* Special counters, in some configurations protected by the
320 * page_table_lock, in other configurations by being atomic.
321 */
4294621f 322 mm_counter_t _file_rss;
1da177e4
LT
323 mm_counter_t _anon_rss;
324
f449952b
HD
325 unsigned long hiwater_rss; /* High-watermark of RSS usage */
326 unsigned long hiwater_vm; /* High-water virtual memory usage */
327
328 unsigned long total_vm, locked_vm, shared_vm, exec_vm;
329 unsigned long stack_vm, reserved_vm, def_flags, nr_ptes;
330 unsigned long start_code, end_code, start_data, end_data;
331 unsigned long start_brk, brk, start_stack;
332 unsigned long arg_start, arg_end, env_start, env_end;
333
36d57ac4 334 unsigned long saved_auxv[AT_VECTOR_SIZE]; /* for /proc/PID/auxv */
1da177e4 335
d6e71144 336 unsigned dumpable:2;
1da177e4
LT
337 cpumask_t cpu_vm_mask;
338
339 /* Architecture-specific MM context */
340 mm_context_t context;
341
342 /* Token based thrashing protection. */
343 unsigned long swap_token_time;
344 char recent_pagein;
345
346 /* coredumping support */
347 int core_waiters;
348 struct completion *core_startup_done, core_done;
349
350 /* aio bits */
351 rwlock_t ioctx_list_lock;
352 struct kioctx *ioctx_list;
1da177e4
LT
353};
354
355struct sighand_struct {
356 atomic_t count;
357 struct k_sigaction action[_NSIG];
358 spinlock_t siglock;
359};
360
0e464814
KK
361struct pacct_struct {
362 unsigned long ac_mem;
363};
364
1da177e4
LT
365/*
366 * NOTE! "signal_struct" does not have it's own
367 * locking, because a shared signal_struct always
368 * implies a shared sighand_struct, so locking
369 * sighand_struct is always a proper superset of
370 * the locking of signal_struct.
371 */
372struct signal_struct {
373 atomic_t count;
374 atomic_t live;
375
376 wait_queue_head_t wait_chldexit; /* for wait4() */
377
378 /* current thread group signal load-balancing target: */
379 task_t *curr_target;
380
381 /* shared signal handling: */
382 struct sigpending shared_pending;
383
384 /* thread group exit support */
385 int group_exit_code;
386 /* overloaded:
387 * - notify group_exit_task when ->count is equal to notify_count
388 * - everyone except group_exit_task is stopped during signal delivery
389 * of fatal signals, group_exit_task processes the signal.
390 */
391 struct task_struct *group_exit_task;
392 int notify_count;
393
394 /* thread group stop support, overloads group_exit_code too */
395 int group_stop_count;
396 unsigned int flags; /* see SIGNAL_* flags below */
397
398 /* POSIX.1b Interval Timers */
399 struct list_head posix_timers;
400
401 /* ITIMER_REAL timer for the process */
2ff678b8 402 struct hrtimer real_timer;
05cfb614 403 struct task_struct *tsk;
2ff678b8 404 ktime_t it_real_incr;
1da177e4
LT
405
406 /* ITIMER_PROF and ITIMER_VIRTUAL timers for the process */
407 cputime_t it_prof_expires, it_virt_expires;
408 cputime_t it_prof_incr, it_virt_incr;
409
410 /* job control IDs */
411 pid_t pgrp;
412 pid_t tty_old_pgrp;
413 pid_t session;
414 /* boolean value for session group leader */
415 int leader;
416
417 struct tty_struct *tty; /* NULL if no tty */
418
419 /*
420 * Cumulative resource counters for dead threads in the group,
421 * and for reaped dead child processes forked by this group.
422 * Live threads maintain their own counters and add to these
423 * in __exit_signal, except for the group leader.
424 */
425 cputime_t utime, stime, cutime, cstime;
426 unsigned long nvcsw, nivcsw, cnvcsw, cnivcsw;
427 unsigned long min_flt, maj_flt, cmin_flt, cmaj_flt;
428
429 /*
430 * Cumulative ns of scheduled CPU time for dead threads in the
431 * group, not including a zombie group leader. (This only differs
432 * from jiffies_to_ns(utime + stime) if sched_clock uses something
433 * other than jiffies.)
434 */
435 unsigned long long sched_time;
436
437 /*
438 * We don't bother to synchronize most readers of this at all,
439 * because there is no reader checking a limit that actually needs
440 * to get both rlim_cur and rlim_max atomically, and either one
441 * alone is a single word that can safely be read normally.
442 * getrlimit/setrlimit use task_lock(current->group_leader) to
443 * protect this instead of the siglock, because they really
444 * have no need to disable irqs.
445 */
446 struct rlimit rlim[RLIM_NLIMITS];
447
448 struct list_head cpu_timers[3];
449
450 /* keep the process-shared keyrings here so that they do the right
451 * thing in threads created with CLONE_THREAD */
452#ifdef CONFIG_KEYS
453 struct key *session_keyring; /* keyring inherited over fork */
454 struct key *process_keyring; /* keyring private to this process */
455#endif
0e464814
KK
456#ifdef CONFIG_BSD_PROCESS_ACCT
457 struct pacct_struct pacct; /* per-process accounting information */
458#endif
1da177e4
LT
459};
460
4866cde0
NP
461/* Context switch must be unlocked if interrupts are to be enabled */
462#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
463# define __ARCH_WANT_UNLOCKED_CTXSW
464#endif
465
1da177e4
LT
466/*
467 * Bits in flags field of signal_struct.
468 */
469#define SIGNAL_STOP_STOPPED 0x00000001 /* job control stop in effect */
470#define SIGNAL_STOP_DEQUEUED 0x00000002 /* stop signal dequeued */
471#define SIGNAL_STOP_CONTINUED 0x00000004 /* SIGCONT since WCONTINUED reap */
472#define SIGNAL_GROUP_EXIT 0x00000008 /* group exit in progress */
473
474
475/*
476 * Priority of a process goes from 0..MAX_PRIO-1, valid RT
b0a9499c
IM
477 * priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
478 * tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
479 * values are inverted: lower p->prio value means higher priority.
1da177e4
LT
480 *
481 * The MAX_USER_RT_PRIO value allows the actual maximum
482 * RT priority to be separate from the value exported to
483 * user-space. This allows kernel threads to set their
484 * priority to a value higher than any user task. Note:
485 * MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
486 */
487
488#define MAX_USER_RT_PRIO 100
489#define MAX_RT_PRIO MAX_USER_RT_PRIO
490
491#define MAX_PRIO (MAX_RT_PRIO + 40)
492
493#define rt_task(p) (unlikely((p)->prio < MAX_RT_PRIO))
d425b274 494#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH))
1da177e4
LT
495
496/*
497 * Some day this will be a full-fledged user tracking system..
498 */
499struct user_struct {
500 atomic_t __count; /* reference count */
501 atomic_t processes; /* How many processes does this user have? */
502 atomic_t files; /* How many open files does this user have? */
503 atomic_t sigpending; /* How many pending signals does this user have? */
2d9048e2 504#ifdef CONFIG_INOTIFY_USER
0eeca283
RL
505 atomic_t inotify_watches; /* How many inotify watches does this user have? */
506 atomic_t inotify_devs; /* How many inotify devs does this user have opened? */
507#endif
1da177e4
LT
508 /* protected by mq_lock */
509 unsigned long mq_bytes; /* How many bytes can be allocated to mqueue? */
510 unsigned long locked_shm; /* How many pages of mlocked shm ? */
511
512#ifdef CONFIG_KEYS
513 struct key *uid_keyring; /* UID specific keyring */
514 struct key *session_keyring; /* UID's default session keyring */
515#endif
516
517 /* Hash table maintenance information */
518 struct list_head uidhash_list;
519 uid_t uid;
520};
521
522extern struct user_struct *find_user(uid_t);
523
524extern struct user_struct root_user;
525#define INIT_USER (&root_user)
526
527typedef struct prio_array prio_array_t;
528struct backing_dev_info;
529struct reclaim_state;
530
531#ifdef CONFIG_SCHEDSTATS
532struct sched_info {
533 /* cumulative counters */
534 unsigned long cpu_time, /* time spent on the cpu */
535 run_delay, /* time spent waiting on a runqueue */
536 pcnt; /* # of timeslices run on this cpu */
537
538 /* timestamps */
539 unsigned long last_arrival, /* when we last ran on a cpu */
540 last_queued; /* when we were last queued to run */
541};
542
543extern struct file_operations proc_schedstat_operations;
544#endif
545
546enum idle_type
547{
548 SCHED_IDLE,
549 NOT_IDLE,
550 NEWLY_IDLE,
551 MAX_IDLE_TYPES
552};
553
554/*
555 * sched-domains (multiprocessor balancing) declarations:
556 */
557#ifdef CONFIG_SMP
558#define SCHED_LOAD_SCALE 128UL /* increase resolution of load */
559
560#define SD_LOAD_BALANCE 1 /* Do load balancing on this domain. */
561#define SD_BALANCE_NEWIDLE 2 /* Balance when about to become idle */
562#define SD_BALANCE_EXEC 4 /* Balance on exec */
147cbb4b
NP
563#define SD_BALANCE_FORK 8 /* Balance on fork, clone */
564#define SD_WAKE_IDLE 16 /* Wake to idle CPU on task wakeup */
565#define SD_WAKE_AFFINE 32 /* Wake task to waking CPU */
566#define SD_WAKE_BALANCE 64 /* Perform balancing at task wakeup */
567#define SD_SHARE_CPUPOWER 128 /* Domain members share cpu power */
1da177e4
LT
568
569struct sched_group {
570 struct sched_group *next; /* Must be a circular list */
571 cpumask_t cpumask;
572
573 /*
574 * CPU power of this group, SCHED_LOAD_SCALE being max power for a
575 * single CPU. This is read only (except for setup, hotplug CPU).
576 */
577 unsigned long cpu_power;
578};
579
580struct sched_domain {
581 /* These fields must be setup */
582 struct sched_domain *parent; /* top domain must be null terminated */
583 struct sched_group *groups; /* the balancing groups of the domain */
584 cpumask_t span; /* span of all CPUs in this domain */
585 unsigned long min_interval; /* Minimum balance interval ms */
586 unsigned long max_interval; /* Maximum balance interval ms */
587 unsigned int busy_factor; /* less balancing by factor if busy */
588 unsigned int imbalance_pct; /* No balance until over watermark */
589 unsigned long long cache_hot_time; /* Task considered cache hot (ns) */
590 unsigned int cache_nice_tries; /* Leave cache hot tasks for # tries */
591 unsigned int per_cpu_gain; /* CPU % gained by adding domain cpus */
7897986b
NP
592 unsigned int busy_idx;
593 unsigned int idle_idx;
594 unsigned int newidle_idx;
595 unsigned int wake_idx;
147cbb4b 596 unsigned int forkexec_idx;
1da177e4
LT
597 int flags; /* See SD_* */
598
599 /* Runtime fields. */
600 unsigned long last_balance; /* init to jiffies. units in jiffies */
601 unsigned int balance_interval; /* initialise to 1. units in ms. */
602 unsigned int nr_balance_failed; /* initialise to 0 */
603
604#ifdef CONFIG_SCHEDSTATS
605 /* load_balance() stats */
606 unsigned long lb_cnt[MAX_IDLE_TYPES];
607 unsigned long lb_failed[MAX_IDLE_TYPES];
608 unsigned long lb_balanced[MAX_IDLE_TYPES];
609 unsigned long lb_imbalance[MAX_IDLE_TYPES];
610 unsigned long lb_gained[MAX_IDLE_TYPES];
611 unsigned long lb_hot_gained[MAX_IDLE_TYPES];
612 unsigned long lb_nobusyg[MAX_IDLE_TYPES];
613 unsigned long lb_nobusyq[MAX_IDLE_TYPES];
614
615 /* Active load balancing */
616 unsigned long alb_cnt;
617 unsigned long alb_failed;
618 unsigned long alb_pushed;
619
68767a0a
NP
620 /* SD_BALANCE_EXEC stats */
621 unsigned long sbe_cnt;
622 unsigned long sbe_balanced;
1da177e4
LT
623 unsigned long sbe_pushed;
624
68767a0a
NP
625 /* SD_BALANCE_FORK stats */
626 unsigned long sbf_cnt;
627 unsigned long sbf_balanced;
628 unsigned long sbf_pushed;
629
1da177e4
LT
630 /* try_to_wake_up() stats */
631 unsigned long ttwu_wake_remote;
632 unsigned long ttwu_move_affine;
633 unsigned long ttwu_move_balance;
634#endif
635};
636
1a20ff27
DG
637extern void partition_sched_domains(cpumask_t *partition1,
638 cpumask_t *partition2);
198e2f18 639
640/*
641 * Maximum cache size the migration-costs auto-tuning code will
642 * search from:
643 */
644extern unsigned int max_cache_size;
645
646#endif /* CONFIG_SMP */
1da177e4
LT
647
648
649struct io_context; /* See blkdev.h */
650void exit_io_context(void);
651struct cpuset;
652
653#define NGROUPS_SMALL 32
654#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
655struct group_info {
656 int ngroups;
657 atomic_t usage;
658 gid_t small_block[NGROUPS_SMALL];
659 int nblocks;
660 gid_t *blocks[0];
661};
662
663/*
664 * get_group_info() must be called with the owning task locked (via task_lock())
665 * when task != current. The reason being that the vast majority of callers are
666 * looking at current->group_info, which can not be changed except by the
667 * current task. Changing current->group_info requires the task lock, too.
668 */
669#define get_group_info(group_info) do { \
670 atomic_inc(&(group_info)->usage); \
671} while (0)
672
673#define put_group_info(group_info) do { \
674 if (atomic_dec_and_test(&(group_info)->usage)) \
675 groups_free(group_info); \
676} while (0)
677
3e30148c
DH
678extern struct group_info *groups_alloc(int gidsetsize);
679extern void groups_free(struct group_info *group_info);
680extern int set_current_groups(struct group_info *group_info);
681extern int groups_search(struct group_info *group_info, gid_t grp);
1da177e4
LT
682/* access the groups "array" with this macro */
683#define GROUP_AT(gi, i) \
684 ((gi)->blocks[(i)/NGROUPS_PER_BLOCK][(i)%NGROUPS_PER_BLOCK])
685
383f2835
CK
686#ifdef ARCH_HAS_PREFETCH_SWITCH_STACK
687extern void prefetch_stack(struct task_struct*);
688#else
689static inline void prefetch_stack(struct task_struct *t) { }
690#endif
1da177e4
LT
691
692struct audit_context; /* See audit.c */
693struct mempolicy;
b92ce558 694struct pipe_inode_info;
1da177e4 695
3dee386e
CK
696enum sleep_type {
697 SLEEP_NORMAL,
698 SLEEP_NONINTERACTIVE,
699 SLEEP_INTERACTIVE,
700 SLEEP_INTERRUPTED,
701};
702
1da177e4
LT
703struct task_struct {
704 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
705 struct thread_info *thread_info;
706 atomic_t usage;
707 unsigned long flags; /* per process flags, defined below */
708 unsigned long ptrace;
709
36772092 710 int lock_depth; /* BKL lock depth */
1da177e4 711
d6077cb8 712#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
4866cde0
NP
713 int oncpu;
714#endif
1da177e4
LT
715 int prio, static_prio;
716 struct list_head run_list;
717 prio_array_t *array;
718
22e2c507 719 unsigned short ioprio;
2056a782 720 unsigned int btrace_seq;
22e2c507 721
1da177e4
LT
722 unsigned long sleep_avg;
723 unsigned long long timestamp, last_ran;
724 unsigned long long sched_time; /* sched_clock time spent running */
3dee386e 725 enum sleep_type sleep_type;
1da177e4
LT
726
727 unsigned long policy;
728 cpumask_t cpus_allowed;
729 unsigned int time_slice, first_time_slice;
730
731#ifdef CONFIG_SCHEDSTATS
732 struct sched_info sched_info;
733#endif
734
735 struct list_head tasks;
736 /*
737 * ptrace_list/ptrace_children forms the list of my children
738 * that were stolen by a ptracer.
739 */
740 struct list_head ptrace_children;
741 struct list_head ptrace_list;
742
743 struct mm_struct *mm, *active_mm;
744
745/* task state */
746 struct linux_binfmt *binfmt;
747 long exit_state;
748 int exit_code, exit_signal;
749 int pdeath_signal; /* The signal sent when the parent dies */
750 /* ??? */
751 unsigned long personality;
752 unsigned did_exec:1;
753 pid_t pid;
754 pid_t tgid;
755 /*
756 * pointers to (original) parent process, youngest child, younger sibling,
757 * older sibling, respectively. (p->father can be replaced with
758 * p->parent->pid)
759 */
760 struct task_struct *real_parent; /* real parent process (when being debugged) */
761 struct task_struct *parent; /* parent process */
762 /*
763 * children/sibling forms the list of my children plus the
764 * tasks I'm ptracing.
765 */
766 struct list_head children; /* list of my children */
767 struct list_head sibling; /* linkage in my parent's children list */
768 struct task_struct *group_leader; /* threadgroup leader */
769
770 /* PID/PID hash table linkage. */
92476d7f 771 struct pid_link pids[PIDTYPE_MAX];
47e65328 772 struct list_head thread_group;
1da177e4
LT
773
774 struct completion *vfork_done; /* for vfork() */
775 int __user *set_child_tid; /* CLONE_CHILD_SETTID */
776 int __user *clear_child_tid; /* CLONE_CHILD_CLEARTID */
777
778 unsigned long rt_priority;
779 cputime_t utime, stime;
780 unsigned long nvcsw, nivcsw; /* context switch counts */
781 struct timespec start_time;
782/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
783 unsigned long min_flt, maj_flt;
784
785 cputime_t it_prof_expires, it_virt_expires;
786 unsigned long long it_sched_expires;
787 struct list_head cpu_timers[3];
788
789/* process credentials */
790 uid_t uid,euid,suid,fsuid;
791 gid_t gid,egid,sgid,fsgid;
792 struct group_info *group_info;
793 kernel_cap_t cap_effective, cap_inheritable, cap_permitted;
794 unsigned keep_capabilities:1;
795 struct user_struct *user;
796#ifdef CONFIG_KEYS
b5f545c8 797 struct key *request_key_auth; /* assumed request_key authority */
1da177e4 798 struct key *thread_keyring; /* keyring private to this thread */
3e30148c 799 unsigned char jit_keyring; /* default keyring to attach requested keys to */
1da177e4
LT
800#endif
801 int oomkilladj; /* OOM kill score adjustment (bit shift). */
36772092
PBG
802 char comm[TASK_COMM_LEN]; /* executable name excluding path
803 - access with [gs]et_task_comm (which lock
804 it with task_lock())
805 - initialized normally by flush_old_exec */
1da177e4
LT
806/* file system info */
807 int link_count, total_link_count;
808/* ipc stuff */
809 struct sysv_sem sysvsem;
810/* CPU-specific state of this task */
811 struct thread_struct thread;
812/* filesystem information */
813 struct fs_struct *fs;
814/* open file information */
815 struct files_struct *files;
816/* namespace */
817 struct namespace *namespace;
818/* signal handlers */
819 struct signal_struct *signal;
820 struct sighand_struct *sighand;
821
822 sigset_t blocked, real_blocked;
150256d8 823 sigset_t saved_sigmask; /* To be restored with TIF_RESTORE_SIGMASK */
1da177e4
LT
824 struct sigpending pending;
825
826 unsigned long sas_ss_sp;
827 size_t sas_ss_size;
828 int (*notifier)(void *priv);
829 void *notifier_data;
830 sigset_t *notifier_mask;
831
832 void *security;
833 struct audit_context *audit_context;
834 seccomp_t seccomp;
835
836/* Thread group tracking */
837 u32 parent_exec_id;
838 u32 self_exec_id;
839/* Protection of (de-)allocation: mm, files, fs, tty, keyrings */
840 spinlock_t alloc_lock;
841/* Protection of proc_dentry: nesting proc_lock, dcache_lock, write_lock_irq(&tasklist_lock); */
842 spinlock_t proc_lock;
1da177e4 843
408894ee
IM
844#ifdef CONFIG_DEBUG_MUTEXES
845 /* mutex deadlock detection */
846 struct mutex_waiter *blocked_on;
847#endif
848
1da177e4
LT
849/* journalling filesystem info */
850 void *journal_info;
851
852/* VM state */
853 struct reclaim_state *reclaim_state;
854
855 struct dentry *proc_dentry;
856 struct backing_dev_info *backing_dev_info;
857
858 struct io_context *io_context;
859
860 unsigned long ptrace_message;
861 siginfo_t *last_siginfo; /* For ptrace use. */
862/*
863 * current io wait handle: wait queue entry to use for io waits
864 * If this thread is processing aio, this points at the waitqueue
865 * inside the currently handled kiocb. It may be NULL (i.e. default
866 * to a stack based synchronous wait) if its doing sync IO.
867 */
868 wait_queue_t *io_wait;
869/* i/o counters(bytes read/written, #syscalls */
870 u64 rchar, wchar, syscr, syscw;
871#if defined(CONFIG_BSD_PROCESS_ACCT)
872 u64 acct_rss_mem1; /* accumulated rss usage */
873 u64 acct_vm_mem1; /* accumulated virtual memory usage */
874 clock_t acct_stimexpd; /* clock_t-converted stime since last update */
875#endif
876#ifdef CONFIG_NUMA
877 struct mempolicy *mempolicy;
878 short il_next;
879#endif
880#ifdef CONFIG_CPUSETS
881 struct cpuset *cpuset;
882 nodemask_t mems_allowed;
883 int cpuset_mems_generation;
825a46af 884 int cpuset_mem_spread_rotor;
1da177e4 885#endif
0771dfef 886 struct robust_list_head __user *robust_list;
34f192c6
IM
887#ifdef CONFIG_COMPAT
888 struct compat_robust_list_head __user *compat_robust_list;
889#endif
0771dfef 890
22e2c507 891 atomic_t fs_excl; /* holding fs exclusive resources */
e56d0903 892 struct rcu_head rcu;
b92ce558
JA
893
894 /*
895 * cache last used pipe for splice
896 */
897 struct pipe_inode_info *splice_pipe;
1da177e4
LT
898};
899
900static inline pid_t process_group(struct task_struct *tsk)
901{
902 return tsk->signal->pgrp;
903}
904
905/**
906 * pid_alive - check that a task structure is not stale
907 * @p: Task structure to be checked.
908 *
909 * Test if a process is not yet dead (at most zombie state)
910 * If pid_alive fails, then pointers within the task structure
911 * can be stale and must not be dereferenced.
912 */
913static inline int pid_alive(struct task_struct *p)
914{
92476d7f 915 return p->pids[PIDTYPE_PID].pid != NULL;
1da177e4
LT
916}
917
918extern void free_task(struct task_struct *tsk);
1da177e4 919#define get_task_struct(tsk) do { atomic_inc(&(tsk)->usage); } while(0)
e56d0903 920
158d9ebd 921extern void __put_task_struct(struct task_struct *t);
e56d0903
IM
922
923static inline void put_task_struct(struct task_struct *t)
924{
925 if (atomic_dec_and_test(&t->usage))
8c7904a0 926 __put_task_struct(t);
e56d0903 927}
1da177e4
LT
928
929/*
930 * Per process flags
931 */
932#define PF_ALIGNWARN 0x00000001 /* Print alignment warning msgs */
933 /* Not implemented yet, only for 486*/
934#define PF_STARTING 0x00000002 /* being created */
935#define PF_EXITING 0x00000004 /* getting shut down */
936#define PF_DEAD 0x00000008 /* Dead */
937#define PF_FORKNOEXEC 0x00000040 /* forked but didn't exec */
938#define PF_SUPERPRIV 0x00000100 /* used super-user privileges */
939#define PF_DUMPCORE 0x00000200 /* dumped core */
940#define PF_SIGNALED 0x00000400 /* killed by a signal */
941#define PF_MEMALLOC 0x00000800 /* Allocating memory */
942#define PF_FLUSHER 0x00001000 /* responsible for disk writeback */
943#define PF_USED_MATH 0x00002000 /* if unset the fpu must be initialized before use */
944#define PF_FREEZE 0x00004000 /* this task is being frozen for suspend now */
945#define PF_NOFREEZE 0x00008000 /* this thread should not be frozen */
946#define PF_FROZEN 0x00010000 /* frozen for system suspend */
947#define PF_FSTRANS 0x00020000 /* inside a filesystem transaction */
948#define PF_KSWAPD 0x00040000 /* I am kswapd */
949#define PF_SWAPOFF 0x00080000 /* I am in swapoff */
950#define PF_LESS_THROTTLE 0x00100000 /* Throttle me less: I clean memory */
b31dc66a
JA
951#define PF_BORROWED_MM 0x00200000 /* I am a kthread doing use_mm */
952#define PF_RANDOMIZE 0x00400000 /* randomize virtual address space */
953#define PF_SWAPWRITE 0x00800000 /* Allowed to write to swap */
954#define PF_SPREAD_PAGE 0x01000000 /* Spread page cache over cpuset */
955#define PF_SPREAD_SLAB 0x02000000 /* Spread some slab caches over cpuset */
c61afb18 956#define PF_MEMPOLICY 0x10000000 /* Non-default NUMA mempolicy */
1da177e4
LT
957
958/*
959 * Only the _current_ task can read/write to tsk->flags, but other
960 * tasks can access tsk->flags in readonly mode for example
961 * with tsk_used_math (like during threaded core dumping).
962 * There is however an exception to this rule during ptrace
963 * or during fork: the ptracer task is allowed to write to the
964 * child->flags of its traced child (same goes for fork, the parent
965 * can write to the child->flags), because we're guaranteed the
966 * child is not running and in turn not changing child->flags
967 * at the same time the parent does it.
968 */
969#define clear_stopped_child_used_math(child) do { (child)->flags &= ~PF_USED_MATH; } while (0)
970#define set_stopped_child_used_math(child) do { (child)->flags |= PF_USED_MATH; } while (0)
971#define clear_used_math() clear_stopped_child_used_math(current)
972#define set_used_math() set_stopped_child_used_math(current)
973#define conditional_stopped_child_used_math(condition, child) \
974 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= (condition) ? PF_USED_MATH : 0; } while (0)
975#define conditional_used_math(condition) \
976 conditional_stopped_child_used_math(condition, current)
977#define copy_to_stopped_child_used_math(child) \
978 do { (child)->flags &= ~PF_USED_MATH, (child)->flags |= current->flags & PF_USED_MATH; } while (0)
979/* NOTE: this will return 0 or PF_USED_MATH, it will never return 1 */
980#define tsk_used_math(p) ((p)->flags & PF_USED_MATH)
981#define used_math() tsk_used_math(current)
982
983#ifdef CONFIG_SMP
984extern int set_cpus_allowed(task_t *p, cpumask_t new_mask);
985#else
986static inline int set_cpus_allowed(task_t *p, cpumask_t new_mask)
987{
4098f991 988 if (!cpu_isset(0, new_mask))
1da177e4
LT
989 return -EINVAL;
990 return 0;
991}
992#endif
993
994extern unsigned long long sched_clock(void);
995extern unsigned long long current_sched_time(const task_t *current_task);
996
997/* sched_exec is called by processes performing an exec */
998#ifdef CONFIG_SMP
999extern void sched_exec(void);
1000#else
1001#define sched_exec() {}
1002#endif
1003
1004#ifdef CONFIG_HOTPLUG_CPU
1005extern void idle_task_exit(void);
1006#else
1007static inline void idle_task_exit(void) {}
1008#endif
1009
1010extern void sched_idle_next(void);
1011extern void set_user_nice(task_t *p, long nice);
1012extern int task_prio(const task_t *p);
1013extern int task_nice(const task_t *p);
e43379f1 1014extern int can_nice(const task_t *p, const int nice);
1da177e4
LT
1015extern int task_curr(const task_t *p);
1016extern int idle_cpu(int cpu);
1017extern int sched_setscheduler(struct task_struct *, int, struct sched_param *);
1018extern task_t *idle_task(int cpu);
a2a97982
KO
1019extern task_t *curr_task(int cpu);
1020extern void set_curr_task(int cpu, task_t *p);
1da177e4
LT
1021
1022void yield(void);
1023
1024/*
1025 * The default (Linux) execution domain.
1026 */
1027extern struct exec_domain default_exec_domain;
1028
1029union thread_union {
1030 struct thread_info thread_info;
1031 unsigned long stack[THREAD_SIZE/sizeof(long)];
1032};
1033
1034#ifndef __HAVE_ARCH_KSTACK_END
1035static inline int kstack_end(void *addr)
1036{
1037 /* Reliable end of stack detection:
1038 * Some APM bios versions misalign the stack
1039 */
1040 return !(((unsigned long)addr+sizeof(void*)-1) & (THREAD_SIZE-sizeof(void*)));
1041}
1042#endif
1043
1044extern union thread_union init_thread_union;
1045extern struct task_struct init_task;
1046
1047extern struct mm_struct init_mm;
1048
1049#define find_task_by_pid(nr) find_task_by_pid_type(PIDTYPE_PID, nr)
1050extern struct task_struct *find_task_by_pid_type(int type, int pid);
1051extern void set_special_pids(pid_t session, pid_t pgrp);
1052extern void __set_special_pids(pid_t session, pid_t pgrp);
1053
1054/* per-UID process charging. */
1055extern struct user_struct * alloc_uid(uid_t);
1056static inline struct user_struct *get_uid(struct user_struct *u)
1057{
1058 atomic_inc(&u->__count);
1059 return u;
1060}
1061extern void free_uid(struct user_struct *);
1062extern void switch_uid(struct user_struct *);
1063
1064#include <asm/current.h>
1065
1066extern void do_timer(struct pt_regs *);
1067
1068extern int FASTCALL(wake_up_state(struct task_struct * tsk, unsigned int state));
1069extern int FASTCALL(wake_up_process(struct task_struct * tsk));
1070extern void FASTCALL(wake_up_new_task(struct task_struct * tsk,
1071 unsigned long clone_flags));
1072#ifdef CONFIG_SMP
1073 extern void kick_process(struct task_struct *tsk);
1074#else
1075 static inline void kick_process(struct task_struct *tsk) { }
1076#endif
476d139c 1077extern void FASTCALL(sched_fork(task_t * p, int clone_flags));
1da177e4
LT
1078extern void FASTCALL(sched_exit(task_t * p));
1079
1080extern int in_group_p(gid_t);
1081extern int in_egroup_p(gid_t);
1082
1083extern void proc_caches_init(void);
1084extern void flush_signals(struct task_struct *);
1085extern void flush_signal_handlers(struct task_struct *, int force_default);
1086extern int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info);
1087
1088static inline int dequeue_signal_lock(struct task_struct *tsk, sigset_t *mask, siginfo_t *info)
1089{
1090 unsigned long flags;
1091 int ret;
1092
1093 spin_lock_irqsave(&tsk->sighand->siglock, flags);
1094 ret = dequeue_signal(tsk, mask, info);
1095 spin_unlock_irqrestore(&tsk->sighand->siglock, flags);
1096
1097 return ret;
1098}
1099
1100extern void block_all_signals(int (*notifier)(void *priv), void *priv,
1101 sigset_t *mask);
1102extern void unblock_all_signals(void);
1103extern void release_task(struct task_struct * p);
1104extern int send_sig_info(int, struct siginfo *, struct task_struct *);
1105extern int send_group_sig_info(int, struct siginfo *, struct task_struct *);
1106extern int force_sigsegv(int, struct task_struct *);
1107extern int force_sig_info(int, struct siginfo *, struct task_struct *);
1108extern int __kill_pg_info(int sig, struct siginfo *info, pid_t pgrp);
1109extern int kill_pg_info(int, struct siginfo *, pid_t);
1110extern int kill_proc_info(int, struct siginfo *, pid_t);
46113830 1111extern int kill_proc_info_as_uid(int, struct siginfo *, pid_t, uid_t, uid_t);
1da177e4
LT
1112extern void do_notify_parent(struct task_struct *, int);
1113extern void force_sig(int, struct task_struct *);
1114extern void force_sig_specific(int, struct task_struct *);
1115extern int send_sig(int, struct task_struct *, int);
1116extern void zap_other_threads(struct task_struct *p);
1117extern int kill_pg(pid_t, int, int);
1da177e4
LT
1118extern int kill_proc(pid_t, int, int);
1119extern struct sigqueue *sigqueue_alloc(void);
1120extern void sigqueue_free(struct sigqueue *);
1121extern int send_sigqueue(int, struct sigqueue *, struct task_struct *);
1122extern int send_group_sigqueue(int, struct sigqueue *, struct task_struct *);
9ac95f2f 1123extern int do_sigaction(int, struct k_sigaction *, struct k_sigaction *);
1da177e4
LT
1124extern int do_sigaltstack(const stack_t __user *, stack_t __user *, unsigned long);
1125
1126/* These can be the second arg to send_sig_info/send_group_sig_info. */
1127#define SEND_SIG_NOINFO ((struct siginfo *) 0)
1128#define SEND_SIG_PRIV ((struct siginfo *) 1)
1129#define SEND_SIG_FORCED ((struct siginfo *) 2)
1130
621d3121
ON
1131static inline int is_si_special(const struct siginfo *info)
1132{
1133 return info <= SEND_SIG_FORCED;
1134}
1135
1da177e4
LT
1136/* True if we are on the alternate signal stack. */
1137
1138static inline int on_sig_stack(unsigned long sp)
1139{
1140 return (sp - current->sas_ss_sp < current->sas_ss_size);
1141}
1142
1143static inline int sas_ss_flags(unsigned long sp)
1144{
1145 return (current->sas_ss_size == 0 ? SS_DISABLE
1146 : on_sig_stack(sp) ? SS_ONSTACK : 0);
1147}
1148
1da177e4
LT
1149/*
1150 * Routines for handling mm_structs
1151 */
1152extern struct mm_struct * mm_alloc(void);
1153
1154/* mmdrop drops the mm and the page tables */
1155extern void FASTCALL(__mmdrop(struct mm_struct *));
1156static inline void mmdrop(struct mm_struct * mm)
1157{
1158 if (atomic_dec_and_test(&mm->mm_count))
1159 __mmdrop(mm);
1160}
1161
1162/* mmput gets rid of the mappings and all user-space */
1163extern void mmput(struct mm_struct *);
1164/* Grab a reference to a task's mm, if it is not already going away */
1165extern struct mm_struct *get_task_mm(struct task_struct *task);
1166/* Remove the current tasks stale references to the old mm_struct */
1167extern void mm_release(struct task_struct *, struct mm_struct *);
1168
1169extern int copy_thread(int, unsigned long, unsigned long, unsigned long, struct task_struct *, struct pt_regs *);
1170extern void flush_thread(void);
1171extern void exit_thread(void);
1172
1da177e4 1173extern void exit_files(struct task_struct *);
6b3934ef 1174extern void __cleanup_signal(struct signal_struct *);
a7e5328a 1175extern void __cleanup_sighand(struct sighand_struct *);
1da177e4
LT
1176extern void exit_itimers(struct signal_struct *);
1177
1178extern NORET_TYPE void do_group_exit(int);
1179
1da177e4
LT
1180extern void daemonize(const char *, ...);
1181extern int allow_signal(int);
1182extern int disallow_signal(int);
1183extern task_t *child_reaper;
1184
1185extern int do_execve(char *, char __user * __user *, char __user * __user *, struct pt_regs *);
1186extern long do_fork(unsigned long, unsigned long, struct pt_regs *, unsigned long, int __user *, int __user *);
1187task_t *fork_idle(int);
1188
1189extern void set_task_comm(struct task_struct *tsk, char *from);
1190extern void get_task_comm(char *to, struct task_struct *tsk);
1191
1192#ifdef CONFIG_SMP
1193extern void wait_task_inactive(task_t * p);
1194#else
1195#define wait_task_inactive(p) do { } while (0)
1196#endif
1197
1198#define remove_parent(p) list_del_init(&(p)->sibling)
8fafabd8 1199#define add_parent(p) list_add_tail(&(p)->sibling,&(p)->parent->children)
1da177e4 1200
5e85d4ab 1201#define next_task(p) list_entry(rcu_dereference((p)->tasks.next), struct task_struct, tasks)
1da177e4
LT
1202
1203#define for_each_process(p) \
1204 for (p = &init_task ; (p = next_task(p)) != &init_task ; )
1205
1206/*
1207 * Careful: do_each_thread/while_each_thread is a double loop so
1208 * 'break' will not work as expected - use goto instead.
1209 */
1210#define do_each_thread(g, t) \
1211 for (g = t = &init_task ; (g = t = next_task(g)) != &init_task ; ) do
1212
1213#define while_each_thread(g, t) \
1214 while ((t = next_thread(t)) != g)
1215
de12a787
EB
1216/* de_thread depends on thread_group_leader not being a pid based check */
1217#define thread_group_leader(p) (p == p->group_leader)
1da177e4 1218
a9cdf410 1219static inline task_t *next_thread(const task_t *p)
47e65328
ON
1220{
1221 return list_entry(rcu_dereference(p->thread_group.next),
1222 task_t, thread_group);
1223}
1224
1da177e4
LT
1225static inline int thread_group_empty(task_t *p)
1226{
47e65328 1227 return list_empty(&p->thread_group);
1da177e4
LT
1228}
1229
1230#define delay_group_leader(p) \
1231 (thread_group_leader(p) && !thread_group_empty(p))
1232
1da177e4 1233/*
260ea101 1234 * Protects ->fs, ->files, ->mm, ->group_info, ->comm, keyring
22e2c507 1235 * subscriptions and synchronises with wait4(). Also used in procfs. Also
053199ed 1236 * pins the final release of task.io_context. Also protects ->cpuset.
1da177e4
LT
1237 *
1238 * Nests both inside and outside of read_lock(&tasklist_lock).
1239 * It must not be nested with write_lock_irq(&tasklist_lock),
1240 * neither inside nor outside.
1241 */
1242static inline void task_lock(struct task_struct *p)
1243{
1244 spin_lock(&p->alloc_lock);
1245}
1246
1247static inline void task_unlock(struct task_struct *p)
1248{
1249 spin_unlock(&p->alloc_lock);
1250}
1251
f63ee72e
ON
1252extern struct sighand_struct *lock_task_sighand(struct task_struct *tsk,
1253 unsigned long *flags);
1254
1255static inline void unlock_task_sighand(struct task_struct *tsk,
1256 unsigned long *flags)
1257{
1258 spin_unlock_irqrestore(&tsk->sighand->siglock, *flags);
1259}
1260
f037360f
AV
1261#ifndef __HAVE_THREAD_FUNCTIONS
1262
a1261f54 1263#define task_thread_info(task) (task)->thread_info
9fc65876 1264#define task_stack_page(task) ((void*)((task)->thread_info))
a1261f54 1265
10ebffde
AV
1266static inline void setup_thread_stack(struct task_struct *p, struct task_struct *org)
1267{
1268 *task_thread_info(p) = *task_thread_info(org);
1269 task_thread_info(p)->task = p;
1270}
1271
1272static inline unsigned long *end_of_stack(struct task_struct *p)
1273{
1274 return (unsigned long *)(p->thread_info + 1);
1275}
1276
f037360f
AV
1277#endif
1278
1da177e4
LT
1279/* set thread flags in other task's structures
1280 * - see asm/thread_info.h for TIF_xxxx flags available
1281 */
1282static inline void set_tsk_thread_flag(struct task_struct *tsk, int flag)
1283{
a1261f54 1284 set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1285}
1286
1287static inline void clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1288{
a1261f54 1289 clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1290}
1291
1292static inline int test_and_set_tsk_thread_flag(struct task_struct *tsk, int flag)
1293{
a1261f54 1294 return test_and_set_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1295}
1296
1297static inline int test_and_clear_tsk_thread_flag(struct task_struct *tsk, int flag)
1298{
a1261f54 1299 return test_and_clear_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1300}
1301
1302static inline int test_tsk_thread_flag(struct task_struct *tsk, int flag)
1303{
a1261f54 1304 return test_ti_thread_flag(task_thread_info(tsk), flag);
1da177e4
LT
1305}
1306
1307static inline void set_tsk_need_resched(struct task_struct *tsk)
1308{
1309 set_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1310}
1311
1312static inline void clear_tsk_need_resched(struct task_struct *tsk)
1313{
1314 clear_tsk_thread_flag(tsk,TIF_NEED_RESCHED);
1315}
1316
1317static inline int signal_pending(struct task_struct *p)
1318{
1319 return unlikely(test_tsk_thread_flag(p,TIF_SIGPENDING));
1320}
1321
1322static inline int need_resched(void)
1323{
1324 return unlikely(test_thread_flag(TIF_NEED_RESCHED));
1325}
1326
1327/*
1328 * cond_resched() and cond_resched_lock(): latency reduction via
1329 * explicit rescheduling in places that are safe. The return
1330 * value indicates whether a reschedule was done in fact.
1331 * cond_resched_lock() will drop the spinlock before scheduling,
1332 * cond_resched_softirq() will enable bhs before scheduling.
1333 */
1334extern int cond_resched(void);
1335extern int cond_resched_lock(spinlock_t * lock);
1336extern int cond_resched_softirq(void);
1337
1338/*
1339 * Does a critical section need to be broken due to another
1340 * task waiting?:
1341 */
1342#if defined(CONFIG_PREEMPT) && defined(CONFIG_SMP)
1343# define need_lockbreak(lock) ((lock)->break_lock)
1344#else
1345# define need_lockbreak(lock) 0
1346#endif
1347
1348/*
1349 * Does a critical section need to be broken due to another
1350 * task waiting or preemption being signalled:
1351 */
1352static inline int lock_need_resched(spinlock_t *lock)
1353{
1354 if (need_lockbreak(lock) || need_resched())
1355 return 1;
1356 return 0;
1357}
1358
1359/* Reevaluate whether the task has signals pending delivery.
1360 This is required every time the blocked sigset_t changes.
1361 callers must hold sighand->siglock. */
1362
1363extern FASTCALL(void recalc_sigpending_tsk(struct task_struct *t));
1364extern void recalc_sigpending(void);
1365
1366extern void signal_wake_up(struct task_struct *t, int resume_stopped);
1367
1368/*
1369 * Wrappers for p->thread_info->cpu access. No-op on UP.
1370 */
1371#ifdef CONFIG_SMP
1372
1373static inline unsigned int task_cpu(const struct task_struct *p)
1374{
a1261f54 1375 return task_thread_info(p)->cpu;
1da177e4
LT
1376}
1377
1378static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1379{
a1261f54 1380 task_thread_info(p)->cpu = cpu;
1da177e4
LT
1381}
1382
1383#else
1384
1385static inline unsigned int task_cpu(const struct task_struct *p)
1386{
1387 return 0;
1388}
1389
1390static inline void set_task_cpu(struct task_struct *p, unsigned int cpu)
1391{
1392}
1393
1394#endif /* CONFIG_SMP */
1395
1396#ifdef HAVE_ARCH_PICK_MMAP_LAYOUT
1397extern void arch_pick_mmap_layout(struct mm_struct *mm);
1398#else
1399static inline void arch_pick_mmap_layout(struct mm_struct *mm)
1400{
1401 mm->mmap_base = TASK_UNMAPPED_BASE;
1402 mm->get_unmapped_area = arch_get_unmapped_area;
1403 mm->unmap_area = arch_unmap_area;
1404}
1405#endif
1406
1407extern long sched_setaffinity(pid_t pid, cpumask_t new_mask);
1408extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
1409
1da177e4
LT
1410extern void normalize_rt_tasks(void);
1411
1da177e4 1412#ifdef CONFIG_PM
3e1d1d28
CL
1413/*
1414 * Check if a process has been frozen
1415 */
1416static inline int frozen(struct task_struct *p)
1417{
1418 return p->flags & PF_FROZEN;
1419}
1420
1421/*
1422 * Check if there is a request to freeze a process
1423 */
1424static inline int freezing(struct task_struct *p)
1425{
1426 return p->flags & PF_FREEZE;
1427}
1428
1429/*
1430 * Request that a process be frozen
1431 * FIXME: SMP problem. We may not modify other process' flags!
1432 */
1433static inline void freeze(struct task_struct *p)
1434{
1435 p->flags |= PF_FREEZE;
1436}
1437
1438/*
1439 * Wake up a frozen process
1440 */
1441static inline int thaw_process(struct task_struct *p)
1442{
1443 if (frozen(p)) {
1444 p->flags &= ~PF_FROZEN;
1445 wake_up_process(p);
1446 return 1;
1447 }
1448 return 0;
1449}
1450
1451/*
1452 * freezing is complete, mark process as frozen
1453 */
1454static inline void frozen_process(struct task_struct *p)
1455{
1456 p->flags = (p->flags & ~PF_FREEZE) | PF_FROZEN;
1457}
1458
1459extern void refrigerator(void);
1da177e4
LT
1460extern int freeze_processes(void);
1461extern void thaw_processes(void);
1462
3e1d1d28 1463static inline int try_to_freeze(void)
1da177e4 1464{
3e1d1d28
CL
1465 if (freezing(current)) {
1466 refrigerator();
1da177e4
LT
1467 return 1;
1468 } else
1469 return 0;
1470}
1471#else
3e1d1d28
CL
1472static inline int frozen(struct task_struct *p) { return 0; }
1473static inline int freezing(struct task_struct *p) { return 0; }
1474static inline void freeze(struct task_struct *p) { BUG(); }
1475static inline int thaw_process(struct task_struct *p) { return 1; }
1476static inline void frozen_process(struct task_struct *p) { BUG(); }
1477
1478static inline void refrigerator(void) {}
1da177e4
LT
1479static inline int freeze_processes(void) { BUG(); return 0; }
1480static inline void thaw_processes(void) {}
1481
3e1d1d28
CL
1482static inline int try_to_freeze(void) { return 0; }
1483
1da177e4
LT
1484#endif /* CONFIG_PM */
1485#endif /* __KERNEL__ */
1486
1487#endif